Rancang Bangun Sensor Anemometer Berbiaya Rendah Berbasis Sensor Efek Hall dengan Integrasi IoT

Abstract

Stasiun Cuaca Otomatis (AWS) komersial umumnya mahal, mendorong pengembangan AWS berbiaya rendah berbasis cetak 3D. Namun, anemometer berbiaya rendah saat ini masih banyak menggunakan reed switch mekanis yang rentan aus dan menghasilkan data tidak stabil akibat contact bouncing. Penelitian ini merancang bangun sensor anemometer beriaya rendah berbasis sensor magnetik efek Hall bebas bouncing, terintegrasi dengan platform IoT. Sistem menggunakan mikrokontroler Wemos D1 Mini untuk mengakumulasi putaran rotor yang dideteksi sensor efek Hall KY-003 secara non-kontak, lalu mentransmisikan data ke basis data time-series InfluxDB melalui WiFi untuk divisualisasikan real-time pada dasbor Grafana. Biaya komponen inti hanya Rp232.300, lebih terjangkau dibanding produk komersial sejenis. Pengujian kalibrasi menunjukkan hubungan linear kuat antara RPM dan kecepatan angin aktual (R² = 0,960), dengan MAE 0,657 km/jam, RMSE 0,739 km/jam, dan error relatif rata-rata 6,63% pada rentang operasional normal. Pengujian konektivitas IoT 30 hari mencatatkan uptime 94,39%. Rancang bangun ini terbukti andal, mengatasi kelemahan reed switch, dan berpotensi menjadi instrumen pemantau cuaca jarak jauh yang presisi dan terjangkau.

Commercial Automatic Weather Stations (AWS) are generally expensive, prompting development of low-cost AWS using 3D printing. However, current low-cost anemometers still largely rely on mechanical reed switches, prone to wear and unstable data due to contact bouncing. This study designs a low-cost anemometer based on a solid-state Hall effect magnetic sensor free from bouncing, integrated with an IoT platform. The system uses a Wemos D1 Mini microcontroller to accumulate rotor rotations detected by the KY-003 Hall effect sensor non-contactly, then transmits data to an InfluxDB time-series database via WiFi for real-time visualization on a Grafana dashboard. Core component cost is only Rp232,300, more affordable than similar commercial products. Calibration testing showed a strong linear relationship between RPM and actual wind speed (R² = 0.960), with MAE 0.657 km/h, RMSE 0.739 km/h, and average relative error 6.63% within the normal range. 30-day IoT connectivity testing recorded an uptime of 94.39%. This design is reliable, overcomes reed switch weaknesses, and holds strong potential as a precise, affordable remote weather monitoring instrument.